Economics of antibiotic resistance

Abstract

In recent years bacteria have become increasingly resistant to antibiotics, leading to a decline in the effectiveness of antibiotics in treating infectious disease. The first chapter uses a framework based on an epidemiological model of infection in which antibiotic effectiveness is treated as a non-renewable resource. In the model presented, bacterial resistance (the converse of effectiveness) develops as a result of selective pressure on non-resistant strains due to antibiotic use. When two antibiotics differ in quality, it is optimal to use one antibiotic initially, following which it is optimal to switch to a combination of the two drugs. The optimal proportion of the two antibiotics depends on the difference between the rates at which bacterial resistance to each antibiotic evolves and on the differences in their pharmaceutical costs. The theoretical model is estimated using data from Harborview Medical Center on antibiotic resistance of Pseudomonas aeruginosa to a class of antibiotics known as aminoglycosides. In addition to these results, we present a numerical solution to the optimization problem.The second chapter argues in favor of modifications in patent breadth for antibiotics given the unique role of patents as a policy mechanism to encourage optimal antibiotic use. When the impact of current antibiotic use on future resistance is large, broad patents may be socially optimal when the social benefit from preserving effectiveness outweighs the deadweight loss associated with a monopolistic market structure. This result is in sharp contrast with common wisdom in the literature on patent breadth that broad patents only serve to increase welfare costs associated with oligopoly and are therefore undesirable. Further, the use of antibiotics as growth promoters in poultry and livestock both increases the overall level of resistance, as well as encourages (intertemporally) inefficient antibiotic use in humans, further exacerbating the resistance problem. Finally, broad patents encourage investment in marginal cost reducing innovations that may be socially beneficial.The increasing incidence of hospital-acquired infections and the failure of antibiotics to treat these infections is one of the greatest challenges facing modern medicine. Much of the increase in antibiotic resistance has been blamed on widespread antibiotic use in hospitals. Monthly data on both antibiotic resistance and use is becoming more readily available and presents hospital infection committees with an opportunity to estimate the impact of antibiotic use on the resistance of bacterial flora in their hospital. The third and final chapter uses time-series data on antibiotic use and bacterial resistance from Harborview Medical Center, Seattle to jointly estimate (1) the impact of use of aminoglycosides (a class of antibiotics) on the resistance of Pseudomonas aeruginosa (PSAR) to aminoglycosides, and (2) the response of physicians' antibiotic prescribing patterns in response to increases in bacterial resistance. The data reveals that a 10% increase in the number of patients treated with gentamicin resulted in only a 0.23% increase in the level of PSAR resistance to gentamicin. Similarly, a 10% increase in resistance to gentamicin resulted in a roughly 0.8% decrease in the number of patients treated with that antibiotic. Finally, using data on antibiotic use to forecast future resistance may produce more accurate estimates than exponential smoothing.